CaFlow: Enhancing Long-Term Action Quality Assessment with Causal Counterfactual Flow
By: Ruisheng Han , Kanglei Zhou , Shuang Chen and more
Potential Business Impact:
Helps judges score long sports routines better.
Action Quality Assessment (AQA) predicts fine-grained execution scores from action videos and is widely applied in sports, rehabilitation, and skill evaluation. Long-term AQA, as in figure skating or rhythmic gymnastics, is especially challenging since it requires modeling extended temporal dynamics while remaining robust to contextual confounders. Existing approaches either depend on costly annotations or rely on unidirectional temporal modeling, making them vulnerable to spurious correlations and unstable long-term representations. To this end, we propose CaFlow, a unified framework that integrates counterfactual de-confounding with bidirectional time-conditioned flow. The Causal Counterfactual Regularization (CCR) module disentangles causal and confounding features in a self-supervised manner and enforces causal robustness through counterfactual interventions, while the BiT-Flow module models forward and backward dynamics with a cycle-consistency constraint to produce smoother and more coherent representations. Extensive experiments on multiple long-term AQA benchmarks demonstrate that CaFlow achieves state-of-the-art performance. Code is available at https://github.com/Harrison21/CaFlow
Similar Papers
CounterVQA: Evaluating and Improving Counterfactual Reasoning in Vision-Language Models for Video Understanding
CV and Pattern Recognition
Helps computers imagine "what if" in videos.
Continual Action Quality Assessment via Adaptive Manifold-Aligned Graph Regularization
CV and Pattern Recognition
Helps computers judge actions in videos better.
LeapFactual: Reliable Visual Counterfactual Explanation Using Conditional Flow Matching
Machine Learning (CS)
Shows how to change answers to be correct.